Right Axial Pathway

Cell, Vol. 94, 273–276, August 7, 1998, Copyright 1998 by Cell Press

Links in the Left/Right Axial Pathway Richard P. Harvey Developmental Biology Unit The Victor Chang Cardiac Research Institute Darlinghurst 2010 and University of New South Wales Kensington 2033 Australia The apparently bilaterally symmetrical body plan of vertebrates conceals profound asymmetries of the heart, lungs, visceral organs, vascular system, and brain. Efforts to understand how left/right (L/R) asymmetries arise date back to the last century, and the theoretical problems are deep and challenging. Many fundamental points are still confusing, such as whether, in the course of vertebrate evolution, asymmetry arose in a bilaterally symmetrical ancestor, or whether a L/R-asymmetrical ancestor developed a superficial symmetry (Jefferies et al., 1996). The recent discovery of a number of genes expressed asymmetrically in early vertebrate embryos, before the appearance of morphological asymmetries, marked an important turning point (Levin et al., 1995). A flurry of recent papers, some published in this issue of Cell, considerably extends our understanding of the L/R pathway, and we can now begin to visualize how the chain of L/R information flows from egg to organ. This minireview takes a waltz down that pathway as we currently understand it (Figure 1). While the conservation of handed body asymmetries in all vertebrates suggests that the laterality pathway will also be conserved, information is being gathered from a variety of systems, and to enforce a common pathway onto all vertebrates could be presumptuous. A case in point seems to be the essential role played by sonic hedgehog (Shh) in the chick, but not mouse (Paga´n-Westphal and Tabin, 1998). Thus, some local events may be solved differently in different species (see Figure 1). In the Beginning The point of origin of the L/R pathway is unknown. Examination of a subset of chick conjoined twin embryos, those that are oriented in a head-to-head fashion, clearly demonstrates that a prepattern in the egg does not dictate laterality and that each axis establishes L/R independently (Levin et al., 1997). The normal laterality observed in mouse morula aggregation chimeras suggest that L/R in this species becomes fixed after the morula stage. In frogs, laterality remains flexible to perturbation by signaling molecules until at least the blastula stage (Nascone and Mercola, 1997; Hyatt and Yost, 1998), although a point of flexibility should perhaps not be confused with the point of origin. This is highlighted by an interesting finding. In the first cell cycle of frog development, transient microtubule arrays guide a rotation of cytoplasm relative to cortex, the direction of which defines the orientation of the dorsoventral axis. Treatments that destroy the microtubules prevent both cytoplasmic rotation and dorsalization, although tilting such eggs in the gravitational field can rescue the rotation and therefore the body axis. However, a normal L/R axis is not restored (Yost, 1998). Thus, microtubule

Minireview

arrays aligned in an anteroposterior direction appear to help define the coordinates against which the L/R axial process is oriented. The L/R axis is the third or last body axis to form. As predicted by Brown and Wolpert, a chiral molecule may harness directional coordinates associated with the anteroposterior and dorsoventral axes, for orientation of the L/R pathway (Brown and Wolpert, 1990). In this light it is interesting that disturbances of microtubule-based events also appear to underlie laterality defects in humans and mice. Genes encoding axonemal dyneins, components of microtubule-based molecular motors, are mutated in Kartagener’s syndrome, in which we see a randomized situs inversus (complete reversal of body asymmetry in 50% of affected individuals), and in the mouse iv strain, which shows heterotaxia (discordant reversals of heart and visceral organ situs). Clues to how microtubule-directed events influence situs may come from examination of body asymmetries in C. elegans and snails. Here, the orientation of the mitotic spindle during the first few cleavages appears to be consistently and directionally perturbed, leading to asymmetric cell divisions and downstream asymmetries of lineage and morphology (Wood, 1998). Randomized Asymmetry An unavoidable theoretical issue in this field is the observation of randomized asymmetries (Brown and Wolpert,

Figure 1. Pathway Determining Left/Right Body Asymmetry in Vertebrates Brackets indicate experimental data generated in only one species. Genes in red indicate those known to be expressed asymmetrically in multiple species.

Cell 274

1990). In many cases of disturbed laterality, alterations to organ asymmetry occur randomly across the affected population. For example, in Kartagener’s syndrome, the complete reversal of body asymmetry observed in individuals occurs randomly among carriers of the dynein mutation: one half are reversed (situs inversus), and one half are normal (situs solitus). In some cases, randomization may be due to a block in the flow of laterality information, whereas in others, the flow of information is clearly not blocked but instead confused. In the mouse iv mutation, for example, which shows a heterotaxialike syndrome, genes downstream in the L/R pathway, such as nodal, lefty-2, and Pitx2 (Figure 1), are expressed randomly: sometimes left, sometimes right, sometimes both sides, or not at all (Piedra et al., 1998; Yoshioka et al., 1998 [both in this issue of Cell]). Currently, we can only make guesses as to how such randomizations arise. It would appear that at one or more points in the pathway, distinct left and right identities are established through some sort of L/R tug of war. Upsetting the balance of genes expressed at those times may result in the embryo attempting to reset L/R homeostasis, sometimes successfully, sometimes not (Isaac et al., 1997). It seems plausible that lateral inhibition pathways akin to Notch/Delta, or battles between agonists and antagonists, are involved. The Left-Right Coordinator A framework for thinking about the early steps in the L/R pathway has recently been provided by Yost and colleagues (Hyatt and Yost, 1998) They found that expression of a TGFb superfamily member, Vg1, in a particular cell on the right side of a 16-cell frog embryo (R3), completely reversed heart and gut looping (situs inversus), accompanied by a reversal of nodal expression in lateral plate mesoderm (LPM) from left to right. Vg1 expression in the contralateral L3 cell had no effect. Thus, Vg1 on the right can completely override normal events on the left, consistent with the notion of an early L/R tug of war. When Vg1 was expressed in a different right-sided cell (R1), randomized asymmetry and randomized nodal expression in LPM resulted. Expressed in descendants of this cell, Vg1 may not be able to assert its full repressive influence on the left, leading to a randomized outcome. For the first time, the different forms of situs abnormality found in humans (situs inversus and heterotaxia) can be reproduced experimentally with the same molecule. An important point, however, is that the Vg1 molecule used in these experiments was actually a recombinant prepro-protein called BVg1, designed to mimic native prepro-Vg1 but processed much more efficiently. BVg1 can be processed readily to active Vg1 in embryos, whereas processed native Vg1 has never been detected in vivo and its synthesis is presumably under strict spatial and temporal control. Hyatt and Yost hypothesize that Vg1 is an integral part of a “leftright coordinator,” effectively an organizer of L/R information oriented orthogonal to the dorsoventral axis. The model predicts that native prepro-Vg1 is processed only on the left and that this sets in progress a tug of war that establishes left and right identity. Experiments using signaling antagonists provide a strong argument for the existence of the left-right coordinator, as well as a role

for Vg1. However, more specific antagonists of Vg1 signaling will be required before this concept can be set in stone. The Node as Conduit Asymmetric gene expression in and around Hensen’s node in the chick suggests that this structure is important for setting up and propagating laterality information (Levin et al., 1995; Levin, 1997; Paga´ n-Westphal and Tabin, 1998). A remarkable finding is that the node, and indeed the anterior 40% of the primitive streak, can completely regenerate if ablated, restoring both a full axis and normal L/R (Psychoyos and Stern, 1996). Elegant node rotation and heterochronic grafting experiments by Tabin and colleagues confirm the implication that laterality information is not established autonomously in the node but is imparted to it, probably from lateral tissue (Paga´n-Westphal and Tabin, 1998). The interface with Yost’s left-right coordinator may lie precisely at this point (Figure 1). Once established, asymmetry in the node has a direct bearing on downstream events: rotating a stage 5 node induces nodal expression in right LPM and not, as normally, in left. Thus, laterality information is transmitted to organ progenitors via the node and not directly from perinodal tissue. While the mode of node education remains unresolved, distinct events occurring on the right and left sides of the node most likely bear on the process. Shh, a member of the hedgehog family, appears to be the key signaling intermediate, at least in the chick. Expression is at first bilateral and symmetrical, before becoming asymmetrical and restricted to the left. Leading up to this, two events occur. On the right, both activin bB and Activin receptor IIa are expressed asymmetrically, suggesting that activinbB signaling represses Shh expression on the right (Levin et al., 1995; Levin et al., 1997). This was proven by insertion of activin-soaked beads to the left of the node and follistatin-soaked beads to the right (Isaac et al., 1997; Levin et al., 1997). On the left, Shh expression is maintained and up-regulated, suggesting a direct effect from the left also (Figure 1). That asymmetric Shh is a key intermediate in transfer of laterality information to organ progenitors was shown with a blocking monoclonal antibody (Logan et al., 1998 [this issue of Cell]; Paga´n-Westphal and Tabin, 1998). If antibody-soaked beads were inserted to the left of the node at stage 5, 100% of embryos lacked expression of nodal and Pitx2 in LPM (see Figure 1 and below). A Relay from Shh to LPM The Patched (Ptc) gene encodes a membrane protein that functions in the Shh signaling pathway, and its upregulation is a sensitive barometer of Shh signaling. The limited extent of Ptc expression around the node suggests that Shh is not the signal that directly induces laterality in organ progenitors, which lie out of its range, further to the left. This concept was confirmed in explants and recombinants, and an unknown relay factor “X” appears to transmit the Shh signal leftward (Paga´nWestphal and Tabin, 1998). Morphogenic Inducers The TGFb superfamily members nodal and lefty-2 are good candidates for factors which induce morphological asymmetry in organ primordia. In the mouse they are transiently expressed in very similar domains within

Minireview 275

LPM, including the caudal region of the forming heart tube, where the first cardiac asymmetries are seen, and the visceral primordia. While lefty genes (there are two, lefty-1 and lefty-2) have only been found in mammals, asymmetric nodal expression has been observed in all vertebrate models examined (mouse, chicken, and frog). Furthermore, nodal has been implicated genetically in the laterality pathways of humans and mice (Collignon et al., 1996; Gebbia et al., 1997). Bilateral expression of nodal in chick embryos leads to randomized heart situs and a high proportion of bilaterally symmetrical hearts (Levin et al., 1997). Thus, nodal can influence organ asymmetry, and the randomized organ situs seen with enforced right-sided Shh expression is likely to be a direct consequence of bilateral nodal expression. When nodal and mouse lefty genes were ectopically expressed in the chick on the right side, expression of the Pitx2 gene was induced in LPM (Logan et al., 1998; Yoshioka et al., 1998). Thus, mouse lefty genes are active in the chick, and nodal and lefty-2 appear to have redundant functions. However, their clearly distinct activities in frog embryo assays suggest that a closer scrutiny of their direct and indirect activities is warranted. Transcriptional Regulation of Organ Morphogenesis As reported recently (St. Amand et al., 1998) and in this issue of Cell (Logan et al., 1998; Piedra et al., 1998; Yoshioka et al., 1998), the Pitx2 gene is also expressed asymmetrically in LPM in chick and mouse embryos, and in their asymmetrically developing organs. Pitx2 (also called RIEG/Ptx2/Otlx2/Brx1), which encodes a bicoid-class homeodomain factor, has previously been characterized in the context of pituitary development and as the gene mutated in human Rieger syndrome. In addition to bilateral expression in cephalic mesoderm and other tissues, Pitx2 is expressed in left-sided LPM in a domain very similar to that of nodal (although at a slightly later time), persisting there until well after nodal has been down-regulated in all but the caudal-most cells of the LPM. In the chick, Pitx2 transcripts can then be found along the whole left side of the forming heart tube, with expression continuing in the left side of the atria and ventricles during looping. The pattern was similar in the mouse heart, although perhaps more restricted to its caudal aspect. Pitx2 expression was also seen along the length of the gut tube, including stomach, and continued there during visceral morphogenesis. Functional experiments in chick have been performed with a Pitx2 retrovirus. While expression of nodal and lefty genes in LPM induced Pitx2 (see above), Pitx2 virus could not induce nodal (Logan et al., 1998). Thus, Pitx2 lies downstream of nodal. Pitx2 induced remarkable effects when expressed in heart and visceral progenitors. Infection of right-sided cardiac progenitors, when achieved successfully, induced predominantly bilaterally symmetrical hearts, similar to those induced by bilateral nodal (Levin et al., 1997). These bilateral hearts, which may be left isomerized or abnormal, have never been seen under any other conditions in cultured chick embryos. Left-looped hearts could be generated at increased frequency by first blocking Shh signaling with antibody, then infecting right cardiac progenitors with Pitx2 virus. Reversals of gut looping were also seen after infection of right LPM in the visceral region. Thus, Pitx2,

currently the downstream-most player in the laterality pathway, can regulate heart and visceral morphogenesis. Its expression throughout both cardiac and visceral mesoderm suggests that the laterality pathway delivers the same signal to all asymmetric organs of the body. Each organ, it would seem, interprets the Pitx2 signal according to its own needs. Pitx2 may not do the job alone. A transcription factor gene of the zinc finger class, cSnR, was found to be expressed in right-sided LPM and heart progenitors in the chick (Isaac et al., 1997). Antisense experiments suggest that cSnR functions in the laterality pathway at a time when both nodal and Pitx2 are expressed in left LPM. Blocking cSnR activity caused a randomized reversal of cardiac looping and embryonic turning. It seems that nodal and cSnR can never to be expressed on the same side (Isaac et al., 1997). This is achieved, at least in part, through repression of cSnR by nodal or perhaps Pitx2 (Isaac et al., 1997; Paga´n-Westphal and Tabin, 1998). Thus, cSnR lies downstream of nodal. Since the Drosophila homolog of cSnR (Snail) is a repressor, the function of cSnR may be to inhibit events on the right that would be antagonistic to the pathway on the left. The Midline Barrier A valuable insight into the functioning of the L/R pathway has been contributed by Hamada and colleagues (Meno et al., 1998 [this issue of Cell]). Unlike mouse lefty-2, lefty-1 is normally expressed on the left side of the prospective floorplate of the neural tube, betraying for the first time L/R asymmetry in that structure. The notion of a midline barrier to laterality signals had been formulated previously from embryological and genetic experiments, as well as consideration of laterality defects in conjoined twins. A knockout of lefty-1 now fingers this gene in midline barrier function (Meno et al., 1998). Its deletion results in a variety of laterality defects, interpretable as thoracic left isomerism. In homozygous embryos, expression of left-sided genes such as nodal, lefty-2, and Pitx2 (Figure 1) can be established normally in left LMP, but after a lag, were often seen bilaterally. In the absence of lefty-1, a diffusible signal from the left (perhaps factor X), which normally activates left-sided genes, appears to cross the midline and establish a left identity on the right. Since lefty proteins have properties of BMP inhibitors in frog embryos assays, the major target of lefty-1 may be yet another member of the TGFb superfamily, possibly a BMP. The restriction of isomerisms to the thorax in knockout mice correlates well with the observation that ectopic expression of leftsided genes occurs only anteriorly, perhaps reflecting the normal distribution of the offending diffusible signal (X?). Noji and colleagues have examined the effects of implanting lefty-1– and lefty-2–soaked beads close to the midline of chick embryos at stage 5 (Yoshioka et al. 1998). They find that both proteins can inhibit the laterality pathway (left-sided Pitx2 expression). However, as noted above, lefty proteins also induce Pitx2 in right LPM if beads are implanted at a later stage. It therefore appears that the two proteins have identical properties, but by virtue of the site and timing of their expression, perform very different functions. Early, lefty-1 blocks the

Cell 276

transfer of laterality information across the midline, while later, lefty-2 (along with nodal) induces Pitx2 in organ progenitors. The molecular and evolutionary basis of this economy will be intriguing to dissect. Summary We now have a sketch of the vertebrate L/R pathway from egg to organ. There is a lot to learn and a lot to explain, but the beauty of the pathway is already evident. We also have an inkling of its fragility. Set against the complex laterality disorders seen in humans, efforts to understand this pathway should continue to challenge our intellectual and experimental dexterity. Selected Reading Brown, N.A., and Wolpert, L. (1990). Development 109, 1–9. Collignon, J., Varlet, I., and Robertson, E.J. (1996). Nature 381, 155–158. Gebbia, M., Ferrero, G.B., Pilia, G., Bassi, M.T., Aylsworth, A.S., Penman-Splitt, M., Bird, L.M., Bamforth, J.S., Burn, J., Schlessinger, D., Nelson, D.L., and Casey, B. (1997). Nat. Genet. 17, 305–308. Hyatt, B.A, and Yost, H.J. (1998). Cell 93, 37–46. Isaac, A., Sargent, M.G., and Cooke, J. (1997). Science 275, 1301– 1304. Jefferies, R.P. S., Brown, N.A., and Daley, P.E. J. (1996). Atca Zoologica (Stockholm) 77, 101–122. Levin, M. (1997). Bioessays 19, 287–296. Levin, M., Johnson, R.L., Stern, C.D., Kuehn, M.R., and Tabin, C.J. (1995). Cell 82, 803–814. Levin, M., Pagan, S., Roberts, D.J., Cooke, J., Kuehn, M.R., and Tabin, C.J. (1997). Dev. Biol. 189, 57–67. Logan, M., Paga´n-Westphal, S.M., Smith, D.M., Paganessi, L., and Tabin, C.J. (1998). Cell 94, this issue, 307–317. Meno, C., Shimono, A., Saijoh, Y., Yashiro, K., Mochida, K., Oishi, S., Noji, S., Kondoh, H., and Hamada, H. (1998). Cell 94, this issue, 287–297. Nascone, N., and Mercola, M. (1997). Dev. Biol. 189, 68–78. Paga´n-Westphal, S.M., and Tabin, C.J. (1998). Cell 93, 25–35. Piedra, M.E., Icardo, J.M., Albajar, M., Rodriguez-Rey, J.C., and Ros, M.A. (1998). Cell 94, this issue, 319–324. Psychoyos, D., and Stern, C.D. (1996). Development 122, 3263–3273. St. Amand, T.R., Ra, J., Zhang, Y., Hu, Y., Baber, S.I., Qiu, M., and Chen, Y. (1998). Biochem. Biophys. Res. Comm. 247, 100–105. Wood, W.R. (1998). Semin. Cell Dev. Biol. 9, 53–60. Yoshioka, H., Meno, C., Koshiba, K., Sugihara, M., Itoh, H., Ishimaru, Y., Inoue, T., Ohuchi, H., Semina, E.V., Murray, J.C., Hamada, H., and Noji, S. (1998). Cell 94, this issue, 299–305. Yost, H.J. (1998). Semin. Cell Dev. Biol. 9, 61–66.